INDEX

Explanations

working

np_max-act-logits · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_30.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.30.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

setScale

-0.27

 persist

-0.27

 metric

-0.26

ÅŁ

-0.25

 sÃ¶z

-0.25

inya

-0.25

COOKIE

-0.24

 metrics

-0.24

æī¿è¯º

-0.24

çķĻè¨ĢæĿ¿

-0.24

POSITIVE LOGITS

 Imports

0.27

ãģ¨åħ±

0.26

 drafting

0.25

roducing

0.24

 ratt

0.24

çµĦ

0.24

 Step

0.24

 Paper

0.24

eing

0.24

mob

0.24

Activations Density 0.103%

working

No Comments

No Known Activations